Exchange Rates and Public Opinion on the European Single Currency

Home , Crown (British coin), Crown (currency), Danish krone, Krone, Swedish Krona

Is My Crown Better than Your Euro? Exchange Rates and Public Opinion on the European Single Currency

SARA BINZER HOBOLT

Department of Politics and International Relations University of Oxford [email protected]

PATRICK LEBLOND

Department of International Business HEC Montreal [email protected]

Draft Paper. Comments most welcome. Please do not cite without permission from authors.

Paper presented at the European Union Studies Association (EUSA) Tenth Biennial International Conference, Montreal, Canada * May 17-May 19, 2007

∗ Abstract

The No to the euro in referendums in Denmark and Sweden has been characterized as a public rebellion against an elite project and a sign of a general Euroscepticism among the citizens. However, it is often ignored that support for the euro fluctuates significantly over time in these countries, and hence analysing referendum outcomes simply in terms on static factors will provide only part of the explanation. In contrast to existing studies, this paper provides an analysis of the short-term dynamics in public support for the euro in the period leading up to the referendums. We thus address the question of why public attitudes towards monetary integration vary over time. We argue that at least part of the answer can be found in exchange rate fluctuations. Existing studies have neglected the fact that the national currency is not only a purely monetary indicator, but also carries symbolic weight. The public is therefore less likely to surrender their national currency when it is strong than when it is weak. They are also less willing to accept a replacement currency (e.g. the euro) when it is seen as weak vis-à-vis other world currencies. Our analysis of the two euro campaigns lends credence to our proposition that exchange rates matter. Moreover, we test impact of exchange rate changes on support of the euro using time series analysis. We find that the rapid fall in the value of the euro vis-à-vis the dollar contributed to the Danish rejection of the euro, whereas the strength of the Swedish currency made the Swedes more reluctant to relinquish their crown.

∗ The authors would like to thank Jonathan Beauchesne for his helpful research assistance. We are also grateful to Robert Klemmensen for providing us with Danish government support data.

Introduction

The decline in the value of the euro against the dollar was the single most important reason why we lost the referendum. 1

(Henrik Dam Kristensen, director of the Danish government’s euro referendum campaign)

In a world where the flows of goods, services, capital and, to a lesser extent, people are becoming increasingly global, the leaders of states with small, open economies are questioning themselves as to whether it would not be better economically to adopt a global currency like the US dollar or, now, the euro. In democratic countries, such a decision is unlikely to be taken lightly given the important role that the national currency normally plays as a symbol of people’s identification with the state (Helleiner 2003). Where the government decides that it would be best to replace the national currency with another, more global currency, some form of public consultation will be required in order to legitimise such a politically salient decision. In many cases, this public consultation will take the form of a referendum. This is what happened in Denmark in September 2000 and in Sweden in September 2003. The issue for the government is not only to understand the economic costs and benefits of adopting another currency but also to gain the support of a majority of the population for such a decision. It is, therefore, crucial to be aware of the factors that influence public opinion on monetary integration. Fortunately, there is a small but growing literature on the topic as a result of the introduction of the euro in the European Union (EU) in January 1999. It has focused its attention at both aggregate and individual level determinants of support for the European single currency. However, most of these studies have taken a static view of public opinion on the euro, analysing surveys at a given point in time and, hence, neglecting the dynamics of popular sentiments over time. Static analyses may be best to identify structural (or slow-changing) factors affecting people’s opinion vis-à-vis the euro’s adoption but they cannot take into account those factors that influence the evolution of public opinion on shorter time frames (e.g., over months rather than years). The short-term dynamics of popular support for monetary integration are important for a government

1 Interview by Hobolt with Henrik Dam Kristensen, Copenhagen, January 2004. All interview and newspaper quotes are translated by the authors.

that thinks adopting the euro (or the dollar) is the best decision for the country’s economy. As such, these short-term factors will be determinant in deciding the timing of a popular consultation. One such factor that, surprisingly, has received little attention in the literature is the exchange rate. Most (static) studies speak of the importance of people’s national identity vs. their European one in determining the level of support for the European single currency. However, they neglect the fact that the national currency is also a symbol of the country’s identity. The question, then, is whether the symbolic value that people attach to the national currency is stable or variable over time. If the latter, then we need to understand what causes this fluctuation in value. We would argue that the strength of a currency, as measured by its exchange rate, is the most important measure of its symbolic value. Anecdotal evidence supports this view. In Canada, for example, support for adopting a common currency with the United States increases when the Canadian dollar depreciates vis-à-vis the US dollar and vice versa when it appreciates (Leblond 2003). In Italy, in spite of a certain degree of attachment to their national currency, Italians were quite happy to replace their devalued lira with a potentially strong and stable euro governed by an independent European Central Bank. On the other hand, Germans were reluctant to give up their Deutsche Mark (DM) for the euro since it had come to represent the symbol of Germany’s post-war stability and prosperity (Risse 2003). Thus, the implication is that a currency’s exchange rate –in both the short and long term – should be an important determinant of public opinion vis-à-vis monetary integration. The present paper aims to validate the above-mentioned proposition regarding the importance of the exchange rate as a determinant of public opinion on monetary integration, i.e. replacing the national currency with another currency, whether common (e.g., the euro) or foreign (e.g., the US dollar). For this purpose, it analyses the cases of Denmark and Sweden after the introduction of the euro in January 1999. In opposition to existing studies – which are mostly static single-case or cross- national analyses – our focus is on the short-term dynamics of public opinion support for the euro in Denmark and Sweden, separately. This way, we are able to clearly show how the exchange rate is a key factor in explaining short-term fluctuations in people’s sentiments vis-à-vis their country’s membership in the Eurozone. Interestingly, the role of the exchange rate in influencing Danes’ and Swedes’ opinion

on replacing their crown differs, 2 where the former focus on the euro’s exchange rate with the US dollar whereas the latter focus on the krona’s exchange rate with the euro. This is a result of the two countries adopting different monetary policy and exchange rate regimes (see also Jupille and Leblang 2007). The paper is organised as follows. The first section reviews the existing literature on referendum choices and support for Europe’s single currency. The second section presents our theoretical expectations regarding the relationship between exchange rates and public opinion in Denmark and Sweden. The third section describes the euro referendum campaigns in both Denmark and Sweden in order to show that the exchange rate was indeed a salient issue. The fourth section presents the data and methodology for testing the importance of the exchange rate for public opinion on adopting the euro while the fifth section discusses the results of the statistical analysis. The last section concludes on the importance that the exchange rate plays as a determinant of people’s support for monetary integration.

Explaining referendum choices and support for the euro

Most studies of vote choices in European referendums have focused on the individual- level predictors of voting behaviour, rather than the dynamics of opinion formation over time. These individual-level approaches to voting behaviour in EU referendums can be divided into three schools: the ‘community’ explanation, the ‘second-order election’ school and the ‘utilitarian expectations’ school (see Garry, Marsh and Sinnott 2005; Hobolt 2006). The first school focuses on individuals’ values and beliefs, and argues that voting behaviour in EU referendums reflects people’s underlying broad attitudes towards European integration. This ‘community’ approach, therefore suggests that it is primarily voters’ general fear about loss of sovereignty and national identity in a United States of Europe that drive voting in referendums (Siune et al. 1994a, 1994b; Svensson 1994, 2002). Another competing explanation of voting behaviour in EU referendums is inspired by the ‘second-order’ theory of elections (Reif and Schmitt 1980). The important characteristic of 'second-order' elections (local and regional

2 Denmark’s currency is called the krone while Sweden’s is called the krona . Both names mean crown in English.

elections also fall into this category) is that they are regarded of lesser importance than national elections ('first-order') and, consequently, voter turnout is lower, protest- voting and voter-switching are more common, and national issues tend to dominate the election campaigns. Following this logic, voters are expected to use referendums on European integration as a means of signalling their satisfaction or dissatisfaction with the government. Several studies have applied this second-order model to EU referendums and have linked referendum outcomes with attitudes towards national governments. (Franklin, Marsh and Wlezien 1994; Franklin, Marsh and McLaren 1994; Franklin et al.1995; Franklin 2002; Garry et al. 2005). Finally, a third school contends that utilitarian expectations determine voting behaviour in EU referendums. Matthew Gabel (1998a, 1998b) has explained support for European integration as a function of individuals’ ability to exploit the economic opportunities created by market liberalization in the EU. Hence support for integration should be strongest among those who have the most to gain economically from integration (Gabel 1998a, 1998b; Gabel and Palmer 1995). According to this rational economic actor model, individuals who believe they will benefit economically from European integration will vote yes in an EU referendum, whereas people who believe the opposite will vote No. All of these approaches to referendums can also be found in the literature that emerged after the Danish and Swedish referendums to explain their outcomes.

Explaining euro referendum choices in Denmark and Sweden

Many factors have been identified to explain the Danes’ decision with respect to the euro in September 2000. For instance, Marcussen and Zølner (2003) argue that a majority of the Danish people rejected the euro on the grounds that it was perceived to be an ‘elite’ project, which in an egalitarian society like Denmark is something considered unacceptable. Marcussen (2005: 51-52) adds further that the microeconomic arguments presented by the government in favour of Denmark adopting the euro did not manage to convince a majority of the population to support the euro. In line with the utilitarian approach, those individuals who did believe that the euro would create better conditions for the Danish business community were more likely to vote yes in the referendum (Buch and Hansen 2002). For their part, Jupille and Leblang (2007) find that economic considerations did not play a significant role

in determining voters’ position vis-à-vis the euro. What seems to have been a key consideration for the Danes is what the authors call ‘community’ issues, notably national sovereignty and identity. Following the ‘community approach’, those voters who believed that adopting the euro would mean a loss of sovereignty for Denmark were, not surprisingly, more likely to have voted no in the referendum. The authors also found that voters who had less trust in politicians were also less likely to vote in favour of the euro. Using pre- and post-referendum surveys, de Vreese and Semetko (2004) find that EU scepticism, government disapproval, economic expectations and political ideology were significant factors in explaining voters’ choice on referendum day. The more people were sceptical about the EU, the more likely they were to vote no in the referendum (see also Buch and Hansen 2002). Moreover, in line with the second- order explanation, the more they disapproved of the government (mostly based on people’s feelings about the prime minister), the more likely they were to oppose replacing the krone with the euro. As for economic expectations, people who had pessimistic expectations about their personal economic situation in the near future were less likely to support the euro. Unlike most other studies, de Vreese and Semetko examine the impact of the campaign context on vote choices. They find that the news exposure significantly influences vote choices (see also de Vreese 2004). However, they do not examine the effect of other aspects of the campaign context, such as economic conditions, on opinion formation. As with the Danish case, several authors have tried to explain the outcome of the Swedish referendum on the euro. Like Marcussen and Zølner (2003) for Denmark, Lindahl and Naurin (2005) argue that the cleavage between the general public and the political elite is responsible for Swedes’ rejection of the euro in September 2003 (see also Widfeldt 2004). However, contrary to the situation in Denmark, they find that partisanship (or political ideology) did not matter for the referendum result. Interestingly, though, they say this may be because political elites were split on the euro issue, often within political parties. 3 Aylott (2005) even argues that this division amongst parties that officially advocated joining the euro is one of the main reasons why the no side won. He also argues that the euro-zone economies’ bad performance at the time, especially Germany’s, contributed to convincing a large portion of

3 Obviously, this undermines their argument that a majority of Swedes rejected the euro because it was an elite-driven project.

Swedes that adopting the euro was not a panacea (see also Miles 2004). In addition, the good performance of the Swedish economy at the time convinced a large part of the population that staying outside the euro-zone was unlikely to have dire consequences (Miles and Lindh 2004). In their study, Jupille and Leblang (2007) find that individuals who have higher human capital, as measured by the fact that they are business owners or white collar workers as opposed to blue collar workers or unemployed, are more likely to have voted yes in the referendum. Like in the case of Denmark, the authors find that sovereignty and trust in politicians are important factors in determining a Swede’s vote on the euro question. To summarize, one of the common factors in both Denmark’s and Sweden’s euro referendums identified by the literature is that the euro was an elite-driven project that was rejected by a majority of the population. Preoccupations with sovereignty as well as trust in politicians are also factors that were common to both countries’ referendums, as suggested by the community and second-order explanations. One last point of commonality is that a large portion of people in both countries felt that staying outside the euro-zone was not likely to have dire consequences for their economies, and hence fell less compelled by the general economic sociotropic arguments. These factors, however, have been seen to play significant role in explaining cross-national and individual support for the euro in the more literature on support for monetary integration in Europe.

Explaining support for the euro Both the utilitarian approach and the ‘community’ approach have also been applied specifically to explaining public support for the euro. If economic considerations are an important determinant of general support for European integration, they should matter even more for policies specifically related to economic and monetary integration, it has been argued. Following the rational choice cost-benefit logic, citizens who will gain from increasing trade are likely to be more supportive of the euro, since monetary integration will increase trade dependence and interdependence. Gabel (2001), Banducci et al. (2003), Gabel and Hix (2005), and Jupille and Leblang (2007) find that individuals with high involvement in international trade favour the euro more than individuals employed in the non-tradable sector. Studies of support for the euro have also found that sociotropic economic concerns play a role. For example, Gärtner (1997) finds that citizens in countries with a looser fiscal policy and high

deficits are more likely to support the euro (see also Gabel 2001). Kaltenthaler and Anderson (2001) find that both national economic performance and national identity are factors influencing cross-national variation in support for the common currency. In general, these studies give an insight into cross-national variation in support for the euro, but do not examine the dynamics of public support for the euro over time. An exception is the study by Banducci, Karp and Loedel (2003), which examines public support for the euro from 1990 to 2000, using pooled Eurobarometer survey data. Their multilevel approach allows them to examine the effect of changing economic environment in individual-level support for the euro. They find that citizens are more willing to hand over monetary sovereignty to the European level when their national currency is weak vis-à-vis the dollar. Hence, they find that exchange rates matter, although they do not examine the value of the national currency vis-à-vis the euro, but rather in relation to the dollar. Moreover, they find that as national debt decreases, support for the euro also declines. Building on these studies of voting behaviour in referendums and support for the euro, this paper seeks to examine the particular effect of exchange rates on support for the euro over time. Studying vote choices as a dynamic process rather than a static decision is potentially very important if we want to understand the outcomes of these referendums. The campaign period is considered to have a greater influence on public opinion in referendums compared with elections, since the electoral context differs significantly from national elections (de Vreese and Semetko 2005; Hobolt 2005). Most importantly, referendums are generally characterized by a higher degree of electoral volatility. Referendum issues are often relatively unfamiliar to voters, who therefore do not have firm pre-existing attitudes towards the issue at stake (Franklin 2002; LeDuc 2002). Moreover, referendums are considered to be second-order type of national elections with low salience and low levels of involvement (e.g., Franklin et al. 1994). Party identification generally matters less in referendums compared to national elections, because no party name appear on the ballot and because parties may be internally split on the issue (Butler and Ranney, 1994; Denver 2002; Hobolt 2007). If voters know little about the specific ballot proposal and are relatively unconstrained by predispositions and party loyalties, they are more likely to be influenced by changes in the economic and political context. It has been shown in previous studies that there are often significant shifts in public opinion over the course of a referendum campaign (LeDuc 2002; Magleby 1989). Indeed, if we look at the

five year periods leading up to the Danish and Swedish referendums, we do find significant changes in vote intention. Figure 1 shows the development in Danish vote intention the period leading up to the referendum on joining the common currency. Whereas the number of undecided voters remains relatively stable and low around 10 per cent, there are substantive shifts in the proportion of voters favourable to accession to the euro over the five year period. In the period before the introduction of the euro in 1999, 54 per cent of Danish voters are planning to vote no, but this drops to an average of just 39 per cent in the period after the introduction of the euro. However, aggregate vote intention remains volatile even after 1999. In the case of Sweden, Figure 2 also illustrates a high level of volatility in vote intention. Close to half of voters were in favour of the euro when it was introduced in January 1999. Then support decreased unsteadily until 2001, only to go back up to a majority around the introduction of notes and coins in January 2002. Afterwards and until the referendum in September 2003, the percentage of Swedish voters indicating that they would vote for replacing the krona with the euro dropped to less than 40 per cent.

Figure 1 Development in vote intention in Denmark, 1994-2000 80 60 40 Vote intention % 20 0

01jul1994 01jan1996 01jul1997 01jan1999 01jul2000 Date

yes % no % don't know %

Figure 2 Development in vote intention in Sweden, 1999-2003

60 50 40 30 Vote intention % 20 10 01jan1999 01jul2000 01jan2002 01jul2003 Date

yes % no % Don't Know %

These figures thus reveal a short-term dynamics in the support for the euro in Denmark and Sweden. In a context marked by high volatility in public opinion, it is important to not only examine the determinants of the final vote choices, but also analyse how the context shapes the development in vote intention prior to the vote. However, none of the studies of the two euro campaigns explicitly examine the development in vote intention prior to the vote, nor do most the studies on support for the euro examine cross-temporal variation. This study seeks to fill this gap in the literature by presenting a dynamic model of public support for the euro in the Danish and the Swedish cases, focusing on how exchange rates changes affect changes in opinion. Prior to presenting this model, we explore the campaign environment in Denmark and Sweden and illustrate the saliency of the currency values in the debate. First, we discuss how exchange rates may influence support for monetary integration.

Theoretical expectations: exchange rates and public opinion

In Figures 1 and 2, we can observe that popular support for the euro has fluctuated through time. This means that the timing of the Danish and Swedish referendums

played a crucial role in defining the results that obtained. This is why it is important to understand the factors that determine the dynamics of public opinion on the European single currency. One point in common that the studies reviewed in the previous section have is that they are static, mainly because they are focused on explaining the outcome of a one-time event (i.e. the referendums). Many of the explanatory variables that the authors identify as salient are usually fixed in the short term and change only over longer time periods. For example, rejection of elite-driven projects is something that is likely to be fairly stable through time. The same can be said about EU scepticism, political ideology and trust in politicians. Only two factors identified by the (Danish) studies tend to fluctuate over short periods of time: government approval and economic expectations. Surprisingly, none of the studies of the Danish and Swedish referendums focused their attention on the role of the exchange rate in influencing the referendum outcomes. 4 In the introduction, we indicated that there were good reasons to think that the exchange rate should be an important determinant of public opinion on monetary integration. This is because the national currency is a symbol of national identity. And one way to measure the value of this symbol is via the strength of the national currency. As mentioned in the introduction, there is anecdotal evidence from Canada and Italy that a weaker (i.e. depreciated) currency elicits lower levels of popular attachment. In the case where a currency is strong (i.e. appreciated), as in Germany, the population will tend to be strongly attached to it. For public opinion on monetary integration in general and European Monetary Union (EMU) in particular, this means that people in countries with weaker currencies should be more favourable to the adoption of another currency (e.g., the euro), ceteris paribus , than people in countries with stronger national currencies. Amongst the general studies of public opinion and EMU, only Banducci et al. (2003) include the strength of the exchange rate in their analysis (national currency vis-à-vis the dollar). They find that the stronger a currency is, the less people are willing to abandon it. We can thus formulate our first hypothesis:

4 Jupille and Leblang (2007) do base their analysis on the fact that Denmark and Sweden have different exchange rate regimes. However, they are not concerned with the role of the exchange rate as a determinant of public opinion on the euro.

Hypothesis #1: A stronger national currency should be associated with a lower level of public support for the euro.

Along the same lines, it is reasonable to argue that the currency that one country is thinking of using as a replacement for the national currency has to be strong. After all, who wants to replace one weak currency by another? If a strong national currency has a high symbolic value, a currency of low value can equally be taken as a sign of weakness, both symbolically and economically. This is why the Italians had little qualms about replacing the lira with the euro because they considered the former to be weak whereas they expected the latter to be strong. In the German case, people’s expectations were somewhat opposite. Given that the DM had been the leading currency in Europe for many years, Germans were not convinced that the euro would be as stable and strong as the DM it was meant to replace, even if the former was modelled on the latter. Brettschneider et al. (2003) argue that the depreciated value of the euro vis-à-vis the US dollar between January 1999 and the beginning of 2001 (see Figure 5) and its television coverage explain why Germans had a negative view of the euro at the time. Hence, we would also expect the public to be more reluctant to accept a replacement currency if this is seen to be weak vis-à-vis other world currencies, irrespective of the value of the national currency.

Hypothesis #2: The replacement currency (e.g., the euro) should be a strong currency in order to gain popular support.

Now that we have identified the two hypotheses to be tested in the present study, it is useful to examine the extent to which the exchange rate was salient during the referendum campaigns in Denmark and Sweden.

The referendum campaigns in Denmark and Sweden

The Danish referendum on the euro

The Danes famously sent shockwaves though the European establishment when they rejected the Maastricht Treaty in 1992 by a narrow margin of 50.7 per cent. This verdict was reversed a year later in a second referendum on the Maastricht Treaty, but only

after an agreement had been reached to allow Denmark to opt out of certain policy areas: the single currency (final stage of the EMU), common defense, justice and home affairs and European citizenship. But in the late 1990s, the Danish government was increasingly feeling isolated in Europe due to the Danish opt-outs. This spurred a debate on when to have a referendum to abolish the opt-outs. After a long period where the polls showed a favorable public attitude towards the euro, the centre-left Danish government decided to call a referendum on joining the single currency in 2000. In fact, the Danish Prime Minister Poul Nyrup Rasmussen, who had been one of the key architects behind Denmark’s conditional acceptance of the Maastricht Treaty in 1993, mentioned the possibility of a referendum on the euro in his New Year address in January 1999. He argued that it would be best for Denmark, for employment and the welfare society to join the euro “at some point”. A contributing factor to the timing of the referendum, called in early March 2000 with a referendum date of 28 September 2000, was that the minority government, consisting of the Social Democrats and Social Liberals, hoped that a successful outcome in a referendum would give the government a much needed popularity boost prior to the general election. 5 When the referendum was called, it appeared that this would be a successful strategy. The proposal to join the euro was not only backed by the government, but also by all of the main opposition parties. Only the far-right party, the Danish People’s party, the far-left parties and the small Christian People’s Party were against the proposal. Moreover, the employers associations, most trade unions and 46 of 48 of daily newspapers came out in favour of joining the euro (Downs 2001). Denmark also comfortably met the economic criteria of the EMU and its economy was in cycle with the rest of the euro-zone. What is more, Denmark had reached an agreement with the European Central Bank in 1998 to participate in the new Exchange Rate Mechanism (ERM), which meant that the value of the Danish krone was fixed against a narrow band of the euro. Hence, even outside the euro-zone Denmark had little freedom to follow an independent monetary policy and, in that sense, little was lost economically by joining the euro. Despite the strong position of the yes-side, the campaign became an uphill struggle. During the very long and intensive campaign, the majority in favour of the euro was gradually eroded. The government’s key argument was that the single

5 Interviews with Henrik Dam Kristensen and Niels Helveg Petersen, January 2004.

currency would stimulate economic growth, fuel employment and induce economic stability. They also warned that interest rates would rise if Denmark did not join and that a No vote could lead to the loss of more than 6,000 jobs. In contrast, the no-side focused less on the economic aspects and more on the loss of national sovereignty and the threat of a political union. As the campaign director of the Danish People’s Party, Søren Espersen, has noted:

We ran a positive campaign in favour of the crown. Our slogan was “The Crown and the Fatherland”. We avoided the economic aspects. But that is all the yes-side focused on, and that is why they lost. They focused too much on the economic aspects, and none of it turned out to be true. People no longer believed in them. 6

As Espersen points out, the government’s focus on the economic logic of accession backfired during the campaign. First, the highly reputable Danish Economic Council (“The Wise Men”) published a report in May 2000, which concluded the economic consequences of not joining would be minimal, and that a ‘wait-and-see’ approach was sensible. Second, and perhaps more importantly, the euro’s sustained and steep decline in value created uncertainty about the stability of this currency. The rapid fall in the euro against the US dollar was widely reported in the Danish newspapers and contributed to the feeling that a no-vote may be safer than joining a currency in freefall. 7 The no-side was quick to adopt the argument that it would be risky to adopt such a weak currency. Of course, the Danish currency was equally declining in value against the dollar, as it was pegged to the euro, but this was rarely mentioned in the news coverage. Instead the declining value of the euro was front-page news and subject to heated discussion on the debate pages from the time the euro was launched in 1999. The normally dull topic of exchange rates became one of the most salient issues on the news agenda, and the value of the euro was moved from the business section to the front-page. As most newspapers favoured the introduction of the euro, leader articles would emphasise that a weak currency did not necessarily imply that the euro-project was doomed to fail. Yet, in the minds of voters, the image of the plummeting currency was more powerful. As one of the leader articles commented: ‘The numerous news

6 Interview with Søren Espersen by Hobolt, January 2004. Espersen is Head of Press and Communication for the Danish People’s Party since 1995. He ran the party’s campaigns leading up to the Amsterdam and the Euro referendums. 7 The description of newspaper coverage in this section is based on an analysis of articles in Aktuelt, BT, Berlingske Tidende, Ekstra Bladet, Information, Jyllandsposten, Kristeligt Dagblad, Politiken and Weekendavisen from January 1999 to September 2000.

bulletins about increases in oil prices and the euro in free fall make the Danes take one step back and think, “let’s wait a moment and only join when we know it is safe”’ (Aktuelt 2000). Another leader article, written five months before the referendum, predicted: ‘the weak euro is a serious threat to the government’s ambitions on joining the euro. When the very symbol of the EMU is in such a serious crisis, it will be close to impossible to convince the Danes to vote yes’ (Information 2000). In response to the declining value of the euro and the equally declining levels of support for joining the currency, the Danish Prime Minister attempted to appease voters by asserting that Denmark could join the euro and withdraw at a later stage. However, this argument was rendered ineffective by the President of the Commission, Romano Prodi, who responded that membership of the euro was ‘by definition permanent’. Hence, exchange rates, and particularly the declining value of the euro against the dollar, played a not insignificant role in the Danish referendum campaign. As the Foreign Minister at the time, Niels Helveg Petersen, has noted:

The euro referendum was not well organized by the yes-side. We made a number of mistakes. And the euro was in free fall against the dollar. The core of our argument was that the euro would create stability. The fall of the euro made the no-side’s argument – “Let’s wait and see” – seem very credible. 8

The chief campaign strategist for the government, Henrik Dam Kristensen, has also confirmed this interpretation of events in an interview: ‘The most important factor leading to the decline in public support was the exchange rate between the euro and the dollar. The euro was in free fall. It was impossible for us to explain the connection’.9 Ultimately, faced with the option of choosing to join a declining single currency and relinquishing a symbol of national sovereignty or a adopting a much safer ‘wait and see’ policy, a majority of the Danes chose to the latter. A majority of 53.1 per cent voted no with a turnout of 87.5 per cent.

8 Interview with Niels Helveg Petersen. Helveg Petersen was the Danish Minister of Foreign Affairs from 1993 until 2000, when he resigned after the Danes rejected the euro in a referendum. 9 Interview with Henrik Dam Kristensen. Henrik Dam Kristensen was campaign director of the government’s campaign leading up to the referendum on the Euro in 2000.

The Swedish referendum on the euro The debate about adopting the single currency began soon after Sweden joined the EU in January 1995. As part of its accession agreement, Sweden was bound to enter the EMU once it satisfied all the (Maastricht) criteria. 10 Unlike Denmark and the UK, it did not have an opt-out agreement. Therefore, the Swedish government mandated an expert commission, led by economist Lars Calmfors, to study Sweden’s entry into the euro-zone from an economic and political perspective. The report was published in 1996 (Calmfors et al. 1997). The report concluded that the economic arguments did not favour joining the EMU at the time, though they most probably would in the future. As for the political arguments, they were considered as favouring adopting the single currency. Hence, the government decided to adopt a wait-and-see approach. In October 1997, it presented a bill to the Riksdag that said that Sweden would not take part in the EMU on 1 January 1999, but would wait to see if the economic criteria identified by the Calmfors Commission would be met. Then it would consider asking the Swedish public whether it wanted to replace the krona with the euro. Although it is only in November 2002 that Swedish Prime Minister Göran Persson announced that there would be a referendum on the euro on 14 September 2003, the debate about Sweden being part of the euro-zone was alive and kicking well before that date. It all started with a declaration by the previously sceptical Persson in November 1999 that said that Sweden ‘must eventually join the euro’ (Brown-Humes 1999). One of the main arguments used by Persson and others in favour of the euro is the fact that the krona had a history of weakness, whereby it would be repeatedly devalued by the government to accommodate inflation resulting from a generous welfare system. The depreciation of the krona vis-à-vis the euro that began in the fall of 2000 and continued in 2001 (see Figure 4) only reinforced this point of view (Brown-Humes 2001; George 2001).

The most important, I believe, is that we have got a highly weakened krona. Many Swedes have travelled abroad and they do not think that it is really nice to experience. I believe that it is the main argument for Swedish membership in EMU (comment by Göran Persson, TT 2001). 11

10 The fact that Sweden allowed its currency to float against the euro made it contravene one of the criteria: exchange rate stability. 11 Translation by authors.

The problem with this line of argument for joining the euro-zone is that the exchange rate between the krona and the euro switched direction around January 2002, when euro notes and coins were introduced (see Figure 4). As a result, Prime Minster Person and others in favour of replacing the krona with the euro had to resort to other economic arguments to make their case. For example, being part of the euro-zone would increase Sweden’s trade with the other member-states (see Rose 2000). It would also help decrease interest rates, which would make mortgage payments lower. Furthermore, joining the euro-zone would maintain, if not increase, Sweden’s influence within the EU. The argument was often summed up in terms of Sweden being too small to make it on its own in a globalising world. Opponents to the euro pointed out that Sweden’s economic performance was better than that of the euro-zone, where the economies of France, Germany and Italy were more or less stagnating and where many member-states were in breach or close to be in breach of the Stability and Growth Pact.

The No-side made frequent references to the problem of enforcing the Stability and Growth Pact, to the French and German refusal to abide by its rules, to the high unemployment rate and the low growth rate of the euro area, to the problems of making the EU work honestly and smoothly. The No-side argued that Swedes could not trust the EU to carry out a policy that would be beneficial to Sweden, and that Sweden should therefore maintain its own currency, rejecting the euro (Jonung 2004).

As such, the euro-zone was not the example of economic growth and stability that the yes-side was trying to put forward while Sweden was one of the star performers of the EU. Hence, it was easy to argue that having a flexible exchange rate regime with one’s own national currency was better economically for Sweden. Another argument that became popular amongst Swedes was that a no-vote was not irreversible whereas a yes-vote was. Given the uncertainties associated with the euro-zone economic performance, many people were inclined to ‘wait-and-see’, which they could do by voting no (Brown-Humes 2003). With the Swedish economy doing well, they could afford such an approach. On 14 September 2003, more than 80 per cent of eligible voters took part in the referendum 56.1 per cent voted against the euro while 41.8 per cent voted in favour. It was a bitter defeat for the yes-side, which was considered as the likely winner when the campaign was launched 10 months before. Having decided to run

their campaign on economic arguments, they saw most of them lose their argumentative power during the campaign, first and foremost the weakened krona.

Data and Methods

In this section we outline the data and models used to test the two hypotheses outlined above. Our dependent variable is support for joining the euro. More specifically, we use survey data (monthly average) on the percentage of people who would say yes in a referendum on the euro (as a percentage of all voters, excluding people who say ‘don’t know’). Appendix 1 provides details on the data sources. Figure 3 shows vote intention in Denmark and Sweden from January 1999 to January 2004. We have chosen January 1999 as the starting point for our analysis, since this is the month the euro was introduced. 12 There is no clear trend in either of these time series over this 5- year period, but the two series are correlated at 0.25. Support for joining the euro is generally higher in Denmark than in Sweden. In Denmark, average support is 54.5 per cent with a standard deviation of 3.4. In Sweden, average support is 48.5 per cent with a higher standard deviation of 6.1. When the Danish Prime Minister first mentioned the euro referendum in his New Year’s speech on 1 January 1999, almost 60 per cent of all Danish voters who had made up their minds were in favour of the joining the euro (at the time 50 per cent of voters had decided to vote in favour, 34 per cent against and 16 per cent were still undecided). However, this had dropped to below 50 per cent by June of 2000. Swedish support for the euro also waned after the Danish rejection in September 2000, but increased again from May 2001 until it peaked in January 2002. This increase in support spurred renewed discussions about a referendum on the euro within the Swedish social democratic party. However, support for the euro declined in the period leading up to the referendum.

12 Our choice of starting point means that for Sweden we are analysing the results leading up to the referendum, whereas in Denmark we analyse support both before an after the referendum. Given that the value of the euro was made very salient in the Danish referendum campaign, we have no reason to believe that exchange rates would seize to have an influence on public support after the vote. Hence, we have include the entire five-year period in the analysis presented below, but the results are robust for Denmark are robust when we focus on just the period prior to the referendum.

Figure 3 Vote intention in Denmark and Sweden, 1999-2003

70 60 50 Yes voteYes % intention in 40 30 01jan1999 01jul2000 01jan2002 01jul2003 Date

Swedish voters Danish voters

As described above, we are interested in examining the effect of currency developments on support for the euro. First, we want to examine whether the strength of the national currency, the krone (DKK)/krona (SEK), relative to the euro has an effect on vote intention (hypothesis #1). Second, we are interested in examining whether the strength of the euro relative to the US dollar has an impact on public opinion (hypothesis #2). In both cases, we use monthly averaged exchange rate data (see Appendix 1 for more details).

Figure 4 Crown/Euro exchange rate development 9.5 9 8.5 Crown/Euro 8 7.5

01jan1999 01jul2000 01jan2002 01jul2003 Date

SEK/Euro DKK/Euro

Figure 4 shows the exchange rate between the Swedish and the Danish Crown and the euro. This figure shows very little movement in the DKK/euro exchange rate, because the Danish currency was pegged to the euro after January 1999. Unlike Sweden, Denmark had participated in the ERM of the European Monetary System (EMS), maintaining parity with the German Deutschmark since 1982 and then from 1999 with the euro inside the narrow ±2.25% band of ERM II. 13 The only development in the DKK/euro exchange rate is thus a short blip after the Danish rejection of the Euro in September 2000. We therefore would not expect any effect of this exchange rate on Danish public opinion. In contrast, there is quite substantial movement in the Swedish exchange rate. As described above, we expect that as the krona/euro exchange rate increases (i.e. the krona depreciates relative to the euro), public support for the euro will also increase, since a weaker national currency should find less favour with the population. Whilst the DKK/euro exchange rate should not have any impact on Danish public opinion, we expect the exchange rate between the US dollar and the euro to have a significant effect. As described above, the plummeting value of the euro vis-à- vis the dollar came to symbolize the dangers of joining this new currency and

13 See Iversen and Thygesen (1998) for details on Denmark’s exchange rate policies since the 1970s.

relinquishing the Danish krone. It was also widely reported in the media. Figure 5 shows the development in the USD/euro exchange rate. 14

Figure 5 USD/Euro exchange rate development 1.2 1.1 1 USD/Euro .9 .8 01jan1999 01jul2000 01jan2002 01jul2003 Date

Figure 5 clearly shows that the euro was depreciating in the 20 months leading up to the Danish referendum on the euro in September 2000, whereas it was increasing its value against the dollar in the period leading up to the Swedish referendum three years later. In addition to these exchange rate variables, we also include other economic control variables in the model. First, we include a monthly consumer confidence index, which attempts to gauge consumers' feelings about the current condition of the economy and their expectations about the economy's future direction (see de Vreese and Semetko 2004). Second, we include a measure of unemployment as a percentage of total labour force (see Banducci et al. 2003). It measures the state of the economy in a way that may be easier to understand for people than GDP growth. In order to test the theory that referendums are fundamentally about feelings toward the government, we also include a ‘government support’ variable in our model (see de Vreese and

14 Note that a decrease in the value of the USD/euro ratio implies a weaker (i.e. depreciated) euro against the dollar whereas an increase means a stronger (i.e. appreciated) euro vis-à-vis the dollar.

Semetko 2004). This is measured as the percentage of voters who would vote for the governing party (Social Democrats in Sweden) or coalition of parties (in Denmark) if there were an election tomorrow. Following the second-order national election thesis, we would expect that voters are more likely to be in favour of the euro referendum when they are also supportive of the government and intending to vote for the government. Since our data are time series data, we need to take into account time-series dependencies. Failure to attend to these dependencies is likely to lead to spurious results (Granger and Newbold, 1977; Ostrom, 1978). To avoid these problems, we rely on the Box-Jenkins model building procedure of identification-estimation- diagnosis (Box and Jenkins, 1976). We identify the dynamics of the input series, using a univariate Autoregressive, Integrated, Moving Average (ARIMA) model. Checking for trending, we find that both the Danish and the Swedish public opinion series are stationary. We find, however, that both time series are autoregressive first- order processes. To account for this autocorrelation we include a lagged dependent variable. We thus use of lag of Y to model the dynamics in the data. This also makes substantive sense: public support for the euro in month t is partly determined by public opinion in month t-1. Finally, we perform a Q test and plot the residuals. Both tests confirm that the residuals are white noise. In our final tables, we also report the Durbin-Watson statistics, which indicate that autocorrelation is not present. Our causal argument implies that the exchange rate changes come before public opinion changes. Moreover, we expect it to take some time for economic changes to feed into public opinion. We therefore include a one-month time lag in our model. We can express our model in the following way:

= + β + β + β + β + β + β + ε Yt a 1Yt−1 2CrownEuro t−1 3USDEuro t−1 4CCI t−1 5Unemp t−1 6Govt t−1

where the parameter β1 represents the effect of the lagged dependent variable and β2 to β6 captures the effects of the other (lagged) independent variables on support for the euro, and α is the intercept term. In the next section, we discuss the results of estimating this model.

Results

In Tables 1 and 2, we find the statistical results of our estimated model. Our two hypotheses concerned with the relationship between exchange rates and the public opinion for monetary integration are validated. However, some precisions are required. When looking at the results for Denmark in Table 1, we observe that the coefficient for the exchange rate between the krone and the euro is not statistically significant. Although hypothesis #1 expects this relationship to be significant, we saw in Figure 4 that the exchange rate between the krone and the euro is stable throughout the whole period under study because the krone is pegged to the euro. Therefore, as discussed above, we should not see any correlation between public support for the euro and this exchange rate. This is in line with Jupille and Leblang’s (2007) findings that economic calculations were not significant in determining the outcome of the Danish euro referendum, precisely because adopting the euro meant no real change in monetary policy for Denmark as the exchange rate regime would remain a fixed one (so little benefits for individuals exposed to international markets). In such a context, however, we would expect that the strength of the euro would have a strong influence on people’s opinion vis-à-vis adopting the euro. This is what we find in Table 1. The coefficient for the USD/euro exchange rate is positive and statistically significant at the 95 per cent level. Substantively, it means that a 0.1 increase in the exchange rate (i.e. Americans now have to pay 10 cents US more for one euro) is associated with a 1.37 percentage point increase in the proportion of Danes supporting the euro (model 3). In terms of the control variables, the only one that is statistically significant is consumer confidence. The coefficient’s negative sign suggests that as Danes become more confident in their economic prospects, they are less inclined to replace the krone by the euro. This result is in contrast to the one obtained by de Vreese and Semetko (2004). In contrast to the second-order expectation, government support does not have a significant effect on public support for the euro. This also contradicts the results obtained by de Vreese and Semetko (2004) in their study. Interestingly, when people in the same survey were asked whether they support the government in power, they also tended to be in favour of adopting the euro. However, in considering their aggregate behaviour over a period of almost four years, we find no relation between

aggregate support for the government and aggregate support for adopting the euro. The same situation applies to consumer confidence (or personal economic expectations as de Vreese and Semetko [2004] call this factor), except that in this case we find a negative aggregate relation with Danish public opinion on the euro.

Table 1 Support for the euro in Denmark

Model 1 Model 2 Model 3 Coeff. SE Coeff. SE Coeff. SE

Yes vote t-1 0.35** 0.14 0.18 0.16 0.16 0.16

Crown/Euro t-1 -11.88 9.20 - - -8.84 8.98

USD/Euro t-1 - - 14.66** 6.18 13.71** 6.26

Consumer confidence t-1 -0.32* 0.18 -0.37** 0.18 -0.35* 0.18

Unemployment t-1 0.81 0.66 -0.26 0.79 -0.22 0.79

Government Support t-1 0.10 0.09 0.12 0.08 0.10 0.08 Intercept 115.63 71.35 27.51*** 7.25 95.50 69.44

Adj R Squared 0.40 0.44 0.45 DW statistics 2.18 2.10 2.09 N 57 57 57 *** p<0.01 ** p<0.05 * p<0.10

Table 2 Support for the euro in Sweden

Model 1 Model 2 Model 3 Model 4 Coeff. SE Coeff. SE Coeff. SE Coeff. SE

Yes vote t-1 0.72*** 0.09 0.75*** 0.08 0.57*** 0.09 0.58*** 0.09

Crown/Euro t-1 6.54** 2.82 - - 9.00*** 2.71 6.08** 2.62

USD/Euro t-1 - - -19.55** 9.20 -27.86*** 8.76 - -

Crown/USD t-1 ------2.71** 0.91

Consumer confidence t-1 0.13 0.08 -0.11 0.07 0.07 0.08 0.09 0.08

Unemployment t-1 -0.35 0.76 1.99 1.53 3.94** 1.52 3.49** 1.46

Government Support t-1 0.04 0.15 0.22 0.18 0.36** 0.17 0.34** 0.17 Intercept -46.63* 26.19 15.41** 6.81 -64.08*** 24.71 -88.68*** 28.13 Adj R Squared 0.77 0.77 0.81 0.80 DW statistics 1.79 1.83 1.86 1.86 N 56 56 56 56 *** p<0.01 ** p<0.05 * p<0.10

The results for Sweden differ from those obtained for Denmark. In Table 2, we can see that the coefficient for the krona/euro exchange rate is positive and statistically

significant, whether we include the USD/euro exchange rate or not in the regression. This confirms hypothesis #1. Substantively, the result from model 3 means that public support for the euro in Sweden increases by 9 percentage points when Swedes pay an extra krona for one euro (i.e. the krona depreciates vis-à-vis the euro). What is more surprising in Table 2 is the negative and statistically significant coefficient found for the USD/euro exchange rate. This is in direct opposition to hypothesis #2. However, it makes little sense that Swedes would be more favourable to the euro when it is weaker vis-à-vis the dollar, especially when they are more inclined to keep the krona when it is relatively strong against the euro. One way to explain this odd result is to look at the relationship between the krona/USD exchange rate and public support for the euro. In model 4 in Table 2, we can observe that the coefficient for this variable is both positive and statistically as well as substantively significant, even though we include the krona/euro exchange rate in the regression. This suggests that Swedes also attach some importance to their currency’s strength vis-à-vis the world’s other leading currency. In a sense, if one’s national currency is strong against both the dollar and the euro, then it is surely worth keeping; its symbolic value is high. This is also in line with our findings of exchange rates reported in the Swedish media before the referendum. This result for the krona/USD exchange rate implies that the USD/euro exchange rate result is only a statistical artefact that arises because krona/euro = krona/USD * USD/euro. If the relationship between the krona/USD exchange rate and public opinion for the euro is positive, then by definition the relationship between the USD/euro exchange rate and public support for the euro has to be negative. 15 In sum, we can conclude, albeit tentatively, that in the Swedish case, because the krona fluctuates vis-à-vis other currencies, people’s preoccupation is with the krona’s strength not the euro’s. In the Danish, given that the krone was already tied to the euro, it makes sense to focus on the relative strength of the euro, not the krone’s. Again, this is in line with Jupille and Leblang’s (2007) finding that economic calculations in terms of giving up a flexible exchange rate regime for a fixed one were salient in determining voters’ choice for or against adopting the euro. In terms of the control variables, we find in model 3 that coefficients for both unemployment and government support are statistically significant. In the latter case, a positive coefficient is in line with expectations (see de Vreese and Semetko 2004),

15 The pair-wise correlation coefficient between the two exchange rates is -0.91.

whereby support for the government leads to greater support for the euro. However, this result is not robust across the different models presented in Table 2. In the case of unemployment, the result obtained in model 3 suggests that popular support for the euro increases with unemployment, i.e. when the economic is performing less well. This implies that Swedes would see joining the euro-zone as a way to improve the country’s economic performance, ceteris paribus .

Conclusion

We generally expect voters to be ill-informed and apathetic about complex and dreary matters such as monetary integration. Moreover, we know from previous studies of referendums that voters lack factual knowledge about ballot issues in referendums (Bowler and Donnovan 1998; Hobolt 2007). Hence, it is not surprising that the existing literature on the two Scandinavian euro referendums has failed to acknowledge the role of exchange rate in shaping public opinion. First, we do not expect the average voter to pay attention to, let alone have a good understanding of exchange rates. Second, most previous studies have focused on explaining vote choices, rather than changes in opinion formation, and hence dynamic factors, such as exchange rates, have not been included in the analysis. Yet, our study suggests that the omission of exchange rates from the analyses of the two euro referendums is an oversight. People do not need encyclopaedic knowledge of monetary policy in order to be influenced by fluctuations in the exchange rate. In this paper we have argued that the value of a currency represents a symbolic value to many citizens. In other words, the value of a currency vis-à-vis other currencies is used by citizens as a cue or a heuristics as to its more general worth. Whilst economist might argue that a weak currency is in fact good for exports and growth, this is not the type of calculation made by most citizens. Instead, a numerically strong national currency becomes a symbol of national strength, which citizens are less willing to relinquish. Equally, a weak ‘replacement’ currency, which plummets in value vis-à-vis other currencies, represents instability and frailty in the minds of people. Our study of Danish and Swedish campaigns has shown that the symbolism of exchange rate policy was highly salient. However, the framing of the issue of exchange rates was very different in the two campaigns, due to the differences in

exchange rate regimes and economic conditions. In Denmark, where the krone is pegged to the euro, the focus was on the value of the euro vis-à-vis the dollar. The decline in the value of the euro came to symbolise the general weakness and uncertainty about the EMU project, and gave credence to the no-side’s “wait-and-see” argument. The free fall of the value of the euro thus undermined the argument that the Danes should join the euro for ‘economic reasons’. In Sweden, where the krona floats against the euro and other currencies, the focus was on its value, at least originally. When it was depreciating against the euro, many supporters of Sweden’s participation in the EMU claimed that the krona’s weakness was the main reason why it should be replaced by the euro. However, when the krona began appreciating against the euro, the argument fell flat. In fact, it reinforced the no-side’s position that Sweden was better off economically with its own currency. Our time series analyses of support for the euro from 1999 to 2003 corroborate these stories. In Denmark, the value of euro had significant impact on the likelihood of voting no. The weaker the euro against the dollar, the greater the decline in public support for this replacement currency. Given these result, one could argue that the Danish government (unknowingly) chose the worst possible time to hold a referendum on the euro, just as its value had hit rock bottom. In Sweden, it was the value of the national currency vis-à-vis foreign currencies that appears to have shaped public opinion: the stronger the krona, the lower the levels of support for replacing it with the euro. Interestingly, public opinion in Denmark and Sweden was not strongly correlated with feelings about the national government, as the second-order election approach argue. Nor was it significantly affected by short-term changes in unemployment rates. These findings are important for a number of reasons. First, they illustrate the importance of analysing the dynamics of public opinion, rather than focusing solely on the determinants of the final vote choice. Especially in referendums, public opinion may be very volatile in the period leading up to the referendum. Second, they show that context matters, not only the immediate political context of the campaign, but also changes in economic conditions. Finally, the findings suggest that contrary to our view of the apathetic voter, public opinion may react even to complex economic cues, such as exchange rate changes.

References

Aktuelt (2000) ‘Valutakrise i valgkampen’. Leader article 19 September 2000. Aylott, N. (2005) ‘Lessons Learned, Lessons Forgotten: The Swedish Referendum on EMU of September 2003’, Government and Opposition 40(4): 540-564. Anderson, C.J. (1998). When in doubt, use proxies: Attitudes toward domestic politics and support for European integration, Comparative Political Studies 31:(5) 569– 601 Banducci, S., Karp, J.A. and Loedel, P.H. (2003) ‘The euro, economic interests and multi-level governance: Examining support for the common currency’, European Journal of Political Research 42: 685-703. Bowler, S. and T. Donovan (1998) Demanding Choices: Opinion, Voting and Direct Democracy. Ann Arbor: University of Michigan Press. Box, G.E.P. & G.M. Jenkins. (1976). Time Series Analysis: Forecasting and Control. San Francisco: Holden-Day. Brettschneider, F., Maier, M. and Maier, J. (2003) ‘From D-Mark to Euro: The Impact of Mass Media on Public opinion in Germany’, German Politics 12(2): 45-64. Brown-Humes, C. (1999) ‘Swedish markets welcome euro call’, Financial Times , 16 November. Brown-Humes, C. (2001) ‘Persson moving faster than his party’, Financial Times , 9 November. Brown-Humes, C. (2003) ‘Swedes stay cool on euro as battle to join heats up’, Financial Times , 1 August. Buch, R. and Hansen, K.M. (2002) ‘The Danes and Europe: From EC 1972 to Euro 2000 – Elections, Referendums and Attitudes’, Scandinavian Political Studies 25(1): 1-26. Butler, D. and Ranney, A. (eds) (1994). Referendums around the world: The growing use of direct democracy. Washington, DC: Macmillan. Calmfors, L., Flam, H., Gottfries, N., Jerneck, M., Lindahl, R., Matlary, J.H., Nordh Berntsson, C., Rabinowicz, E. and Vredin, A. (1997) EMU - A Swedish Perspective, Amsterdam: Kluwer Academic Publishers. De Vreese, C.H. and Semetko, H.A. (2004) ‘News matters: Influences on the vote in the Danish 2000 euro referendum campaign’, European Journal of Political Research 43: 699-722. Denver, R. (2002). Voting in the 1997 Scottish and Welsh devolution referendums: Information, interests and opinions. European Journal of Political Research 41: 827–843. Downs, William M. (2001) ‘Denmark’s Referendum on the Euro: The Mouse that Roared…Again’. West European Politics 24(1), 222-226. Franklin, M. (2002) ‘Learning from the Danish case: A comment on Palle Svensson’s critique of the Franklin thesis, in European Journal of Political Research 41, 751-757. Franklin, M., C. van der Eijk and M. Marsh (1995) ‘Referendum outcomes and trust in government: Public support for Europe in the wake of Maastricht’ in West European Politics , 18:3 Franklin, M. and C. Wlezien (1997) ’The Responsive Public: Issue Salience, Policy Changes, and Preferences for European Unification’. Journal of Theoretical Politics , 9, 347-63

Franklin, M., M. Marsh, and C. Wlezien (1994) ‘Attitudes toward Europe and Referendum Votes: A Response to Siune and Svensson’ in Electoral Studies , 13:2, 117-121 Franklin, M., M. Marsh and L. McLaren (1994) ‘Uncorking the Bottle: Popular Opposition to European Unification in the Wake of Maastricht’, Journal of Common Market Studies , 32:4 Gabel, M. J. (1998a) Interest and Integration. Market Liberalization, Public Opinion and European Union, Ann Abor: The University of Michigan Press. Gabel, M. J. (1998b) ‘Public Support for European Integration: An Empirical Test of Five Theories’ in The Journal of Politics , 60:2, 333-54 Gabel, M. J. (2001). Divided opinion, common currency: The political economy of public support for EMU. in B. Eichengreen & J.A. Frieden (eds.), The political economy of European monetary unification, 2nd edn. Boulder, CO:Westview Press. Gabel, M.J. and Hix, S. (2005) ‘Understanding Public Support for British Membership of the Single Currency’, Political Studies 53(1): 65-81. Gabel, M.J. and H.D. Palmer (1995) ‘Understanding variation in public support for European integration’ in European Journal of Political Research, 27, 3-19 Garry, J., M. Marsh and R. Sinnott (2005) ‘‘Second Order’ Versus ‘Issue Voting’ Effects in EU Referendums: Evidence from the Irish Nice Treaty Referendums’ in European Union Politics 6(2): 201-221. Gärtner, M. (1997). Who wants the euro—and why? Economic explanations of public attitudes towards a single European Currency, Public Choice 93(3–4): 487–510. George, N. (2001) ‘Euro puts Swedish Krona in the shadow’, Financial Times , 22 March. Granger, C.W.J. & P. Newbold. (1977). Forecasting Economic Time Series . NY: Academic Press. Helleiner, E. (2003) The Making of National Money: Territorial Currencies in Historical Perspective , Ithaca: Cornell University Press. Hobolt, S. B. (2005) ‘When Europe Matters: The Impact of Political Information on Voting Behavior in EU Referendums’. Journal of Elections, Public Opinion and Parties ,15(1):85-110. Hobolt, S. B. (2006) ‘Direct Democracy and European Integration’. Journal of European Public Policy , 13(1): 153-166. Hobolt, S. B. (2007) 'Taking Cues on Europe? Voter Competence and Party Endorsements in Referendums on European Integration'. European Journal of Political Research, 46(2): 151-182. Information (2000) ‘Eurosion’. Leader article, 8 May 2000. Iversen, T. and Thygesen, N. (1998) ‘Denmark: From External to Internal Adjustment’, in E. Jones, J. Frieden and F. Torres (eds) Joining Europe’s Monetary Club: The Challenges fro Smaller Member States , London: Macmillan Press Ltd: 61-81. Jonung, L. (2004) ‘To be or not to be in the euro? Benefits and costs of monetary unification as perceived by voters in the Swedish euro referendum 2003’, European Economy , no. 205, June. Jupille, J. and Leblang, D. (2007) ‘Voting for Change: Calculation, Community, and Euro Referendums’, International Organization 61(4): forthcoming. Kaltenthaler, K.C. & Anderson, C.J. (2001). Europeans and their money: Explaining public support for the common European currency. European Journal of Political Research 40(2): 139–170.

Leblond, P. (2003) ‘Symbolism vs. economics: the loonie vs. the greenback’, Policy Options 24(10): 65-69. LeDuc, L. (2002). Referendums and elections: How do campaigns differ? In D.M. Farrell and R. Schmitt-Beck (eds), Do political campaigns matter? Campaign effects in elections and referendums. London: Routledge , pp. 145–162. Lindahl, R. and Naurin, D. (2005) ‘Sweden: The Twin Faces of a Euro-Outsider’, Journal of European Integration 27(1): 65-87. Magleby, D. B. (1989) Direct Legislation: Voting on Ballot Propositions in the United States. Baltimore, MD: Johns Hopkins University Press. Marcussen, M. (2005) ‘Denmark and European Monetary Integration: Out But Far From Over’, Journal of European Integration 27(1): 43-63. Marcussen, M. and Zølner, M. (2003) ‘Monetarism and the Masses: Denmark and Economic Integration in Europe’, Cooperation and Conflict 38(2): 101-123. Miles, L. (2004) ‘Sweden: ‘Hitchhiking’ and the Euro Referendum’, Cooperation and Conflict 39(2): 155-165. Miles, L. and Lindh, M. (2004) ‘After the Referendum: A Swedish Euro Paradox?’, Cooperation and Conflict 39(2): 201-206. Ostrom, V. (1978). Time Series Analysis: Regression Techniques . Newbury Park, NJ: Sage Publications. Reif, K. and H. Schmitt (1980) ‘Nine Second Order National Elections: A Conceptual Framework for the Analysis of European Election Results’. European Journal of Political Research 8: 3-44. Risse, T. (2003) ‘The Euro between national and European identity’, Journal of European Public Policy 10(4): 487-505. Rose, A.K. (2000) ‘One market, one money: the effect of common currencies on Trade’, Economic Policy 15(30): 7-46. Schneider, G. & Weitsman, P.A. (1996). The punishment trap: Integration referendums as popularity contests, Comparative Political Studies 28(4): 582– 607 Siune, K., Svensson, P. and Tonsgaard, O. (1992) –det blev et nej, Aarhus: Politica. Siune, K., Svensson, P. and Tonsgaard, O. (1994a) –fra et nej til et ja , Aarhus: Politica. Siune, K., Svensson, P. and Tonsgaard, O. (1994b) ‘The EU: the Danes said “no” in 1992, but “yes” in 1993: how and why?’, Electoral Studies 13(2): 107–16. Svensson, P. (2002) ‘Five Danish referendums on the European Community and European Union: A critical assessment of the Franklin thesis’ in European Journal of Political Research 41: 733-750 Svensson, P. (1994) ‘The Danish Yes to Maastricht and Edinburgh. The EC Referendum of May 1993’ in Scandinavian Political Studies 17(1): 69-82. TT (2001) ‘Svag krona bakom EMU-uppsving tror Persson’, Dagens Nyheter , 9 September. Widfeldt, A. (2004) ‘Elite Collusion and Public Defiance: Sweden’s Euro Referendum in 2003’, West European Politics 27(3): 503-517.

Appendix 1: Data sources

Support for the Euro: Denmark: Gallup, Sonar, Greens, IFKA, Megafon, Vilstrup, GfK, Eurobarometer. Sweden: Demoskop, TEMO, SIFO, Gallup

Support for the Government: Denmark: Gallup Denmark Sweden: Demoskop, TEMO

Exchange rate data: Danmarks Nationalbank (http://www.nationalbanken.dk ) Sveriges Riksbank ( http://www.riksbank.com )

Unemployment data : Statistics Denmark ( http://www.statbank.dk ) Statistiska centralbyran ( http://www.scb.se/AM0401-EN )

Consumer Confidence Index: Statistics Denmark ( http://www.statbank.dk ) Konjunkturinstitutet (National Institute of Economic Research) ( http://www.konj.se/ )